Aiming at the degradation of control system performance caused by model prediction errors due to parameter perturbation and model mismatch during the operation of permanent magnet synchronous motor (PMSM), a deadbeat predictive control method based on the improved sliding mode disturbance observer is proposed. First, aiming at the inherent chattering problem in sliding mode control, an adaptive hybrid reaching law for boundary layer based on sign function is designed to achieve smooth crossing at zero point. Then, fuzzy adaptive reasoning rules are designed to adjust the boundary layer width in real time to reduce the steady-state error, taking into account the rapidity of approaching sliding mode surface and chattering suppression, and the stability of the controller is proved. Finally, the feedforward compensation of the prediction model is carried out by using deadbeat current predictive control and fuzzy adaptive sliding mode disturbance observer to improve the robustness of the parameters. The experimental results show that this method can effectively compensate the parameter disturbance when the motor model parameters are mismatched, and the steady-state error is small and the dynamic performance is good.

To address the problems of large current output harmonic, complex calculation and the need for additional setting of the neutral-point voltage weighting coefficient in conventional model predictive current control strategy for neutral point clamped three-level inverter, a variable extended space vector model predictive current control strategy is proposed. First, based on the deadbeat principle, the voltage vectors are optimized in the rotating 60° coordinate system; then, the duty ratio is dynamically changed according to the magnitude of the amplitude and phase errors between the reference voltage vector and the actual voltage vector, thus synthesizing a variable extended space vector set with reduced candidate voltage vector set to 8; subsequently, by constructing virtual vectors and introducing the neutral point balancing factor, the neutral point voltage weight coefficient is eliminated and has good neutral point voltage balancing ability; finally, the effectiveness of the proposed strategy is validated through experiments.

The model-based permanent magnet synchronous motor (PMSM) sensorless control technique relies on the motor parameters. However, the parameters’ values of PMSM always change in operation. And due to the magnetic saturation effects, the difference between the apparent inductance and the incremental inductance always exists, which makes it hard to obtain the necessary apparent inductance for sensorless control. To overcome this problem, an online apparent inductance identification method based on minimum adjustment algorithm for PMSM sensorless control is proposed. First, the incremental inductance table at different working points is obtained offline. Then, the asymmetric high-frequency signal injection is used to identify the incremental inductance at one working point. The incremental inductance table is adaptively adjusted by the proposed minimum adjustment algorithm. Finally, the apparent inductance is calculated using the adjusted inductance table. The proposed method is embedded into PMSM sensorless control technique to overcome the inductance mismatch problem. The experimental results verify the effectiveness of this method.

Cascaded H-bridge (CHB) inverters require common-mode voltage (CMV) injection for fault-tolerant operation, which enables the inverter to maintain balanced line-to-line voltage after a fault without additional hardware costs. However, it can potentially lead to unbalanced power in a three-phase system, leading to reverse power flow, which is unwanted in applications. To mitigate these side-effects, a method for injecting CMV that combines the third harmonic and an alternating clamping algorithm is proposed to minimize the fundamental common-mode voltage (FCMV). The approach starts by using alternating clamping to determine CMV, thereby increasing the output voltage of the faulty phase and reducing CMV. Subsequently, the third harmonic is injected to expand the linear modulation region of the CMV, thereby increasing the maximum line-to-line voltage limit achievable without CMV injection. This optimization mitigates the limitation that smaller line-to-line voltages require larger CMV injection of the proposed method. The proposed fault-tolerant operation strategy demonstrates simple calculations and a reduced FCMV, enhancing converter performance after a fault. Simulations and experiments confirm the effectiveness of this method.

Discontinuous pulse width modulation (DPWM) is widely used in high-power, high-switching three-phase inverters. With the improvement of the inverter performance requirements, the negative impact of the common-mode voltage generated by DPWM on the system cannot be ignored. In order to predict and suppress common-mode voltages, it is necessary to establish the mathematical analytical models of common-mode voltages. By applying the double Fourier series theory, the analytical models for common-mode voltage of DPWM0, DPWM1, and DPWMmax are established. The calculated spectra are derived from the analytical models. The characteristic patterns of common-mode voltage are derived and several approaches to suppressing common-mode voltage are given through the analyses and comparisons of these three analytical models and calculated spectra. Simulations and experiments are carried out based on an inverter-driven motor system, and the simulated spectra and experimental spectra of the common-mode voltage are consistent with the calculated spectra, which verifies the correctness and validity of the analytical models.

In medium- and high-voltage applications, the traditional isolated gate driver uses dual isolation channels to transfer the drive signal and drive power respectively, which suffers from a long delay, a large size and a high cost. The single isolation channel with signal-power integrated transmission is crucial for achieving the integration, low cost and high performance of power devices and driver systems. A time-division multiplexing transmission architecture and the corresponding method are proposed to overcome the limitations in the existing elementary signal-power integrated transmission methods, such as limited duty cycle, unidirectional transmission and unstable power transfer, providing an efficient and reliable design scheme by focusing on key technologies including the transmission timing, modulation method and implementation circuit. Aimed at the requirement that the drive signal should be transferred with high fidelity in a wide range, the key factors that restrain the drive signal transmission performance are analyzed in depth, and the laws of interactions among drive signals, modulation carriers and band-width are also formulated, so as to support the transmission of milliwatt-level drive signal and watt-level drive power theoretically. Moreover, the shared-link time-division multiplexing technology is used to address challenges in the stable transmission of drive signals and compatible transmission of bidirectional signals under the full-duty cycle range of drive signals during the integrated transmission, and an integrated driver based on an isolated flyback converter is developed, enabling the forward transmission of drive signals and power by means of pulse duty-cycle modulation and asynchronous timing, as well the reverse transmission of fault signals realized by pulse amplitude modulation. The compact prototype (20 mm×17 mm×16 mm) adopted a planar printed circuit board (PCB) isolated transformer, and the duty cycle range of drive signals covered 0%-100%, with a signal frequency up to 100 kHz. As a result, the full-duplex integrated transmission of drive signals, drive power and reverse fault signals was realized.

Aimed at the widespread problems of low applicability and difficult monitoring in SiC metal-oxide-semiconductor field-effect transistor (SiC MOSFET) junction temperature monitoring methods, a junction temperature monitoring method based on turn-on transient drain-source voltage is proposed. Due to the existence of power loop inductance during the turn-on process, the turn-on transient drain-source voltage can reflect the change rate of current, and it is negatively correlated with the junction temperature through theoretical deduction. On this basis, a junction temperature estimation model based on turn-on transient drain-source voltage is established using a double-pulse test. Compared with the actual temperature, the junction temperature obtained by the junction temperature estimation model has higher accuracy, while the selected temperature sensitive electrical parameters show load-independence during the turn-on process. In addition, this method is simple to measure and calibrate without destroying the device itself, so it has a wider range of applications.

In recent years, the development of DC microgrid has put forward requirements of high-gain and high efficiency for high-gain DC-DC converters. Based on the quasi-Z source, a high-gain soft-switching DC-DC converter with integrated voltage doubling unit is proposed. The voltage doubling unit and the coupling inductor improve the voltage Boost capability of the converter. Meanwhile, the energy of leakage inductance of the coupling inductor is recovered through the control of an active clamp structure, so that the switch realizes zero-voltage switching and the diode realizes zero voltage-zero current switching. As a result, the loss of the converter is reduced, and its performance is improved. In addition, the working principle for the converter was studied in detail, and an experimental prototype with rated power of 200 W was fabricated. The prototype was compared with the converters proposed in five references, and the feasibility and superiority of the novel converter were verified.

To overcome the disadvantages of a double-switch cascaded Boost converter (CBC), a new CBC with switched capacitor (CBC-SC) is proposed, which solves the problem of high switch and diode voltage stress in the CBC by combining a switched capacitor cell with the CBC and further improves the voltage gain. Due to the design of double-switch, this converter can be driven by three different strategies and show different performances. When an interleaved control strategy is adopted, the voltage gain of the converter is significantly improved, and its performance in terms of device voltage stress is also superior. The working principles and performances of the converter under the interleaved, synchronous and complementary control strategies are analyzed, and it is compared with other similar converters. Finally, a 100 W experimental prototype was built, and the feasibility of the converter working under three control strategies and the correctness of the theoretical analysis were verified by experimental results.

The circulating current between the sub-module bridge arms of a modular multilevel converter-high voltage direct current (MMC-HVDC) transmission system distorts the bridge arm current, thus affecting the balance of capacitor voltage. According to the circuit theorem and working characteristics of a modular multilevel converter (MMC), the problem of circulating current in the second harmonic component and the multi-frequency even-order harmonic component of the MMC is analyzed. The extraction and suppression of the second harmonic component are realized by using a notch filter and a quasi-proportional resonant (QPR) controller. The virtual impedance is added to suppress the multi-frequency even-order harmonic component of circulating current in the MMC. The above suppression strategy simultaneously suppresses the second harmonic component and multi-frequency even-order harmonic component, and the effect of circulating current suppression is obvious, indicating that it is suitable for the circulating current suppression control of MMC-HVDC. Simulation results show that the QPR control strategy with virtual impedance can effectively suppress fluctuations in the MMC circulating current waveform, MMC bridge arm current waveform and MMC-HVDC AC-side current.